Research questions

Lecture 4

Aidan Combs

Duke University
SOCIOL 333 - Summer Term 1 2023

2023-05-23

Today

• Introduce the project
• Research questions
• Finish installing

Project

Big idea

• In a nutshell: Ask a sociological question; answer it with statistical analysis
• Instructions: under “Project” on the course website

Structure

• 5 components; ~1/week:

  1. Proposal (May 25; May 30)
  2. Descriptive statistics (June 5)
  3. Results (June 15; June 20)
  4. Paper (June 21; June 28)
  5. Presentation (June 22 or 26)

Why two dates?

• Proposal, results, and paper
• Date 1: when we will do in-class peer review. Push your work to GitHub before class that day. This is a chance to get feedback; it is also part of your project grade.
• Date 2: when you will submit your work for grading. Incorporate your peers’ comments and anything else requested. Deadline is 11:59pm.

Component 1: Proposal

• Develop two plausible sociological research questions you could answer with data
• The goal: To choose and refine one question to move forward with
• Instructions: course website -> Project -> Component 1: Proposal

Research questions

What makes a good sociology research question?

• Specific
• Feasible
• Sociological
• In this class: Is quantitative; involves a relationship between at least 2 variables

Specific

• A question, not a topic

• Too broad: College student binge drinking
• Better: How are college students’ binge drinking behaviors affected by the binge drinking behaviors of their friends?
• Better: How are college students’ opinions on binge drinking affected by the binge drinking behaviors of their friends?

Specific

• Implicates specific variables

• Too broad: What drives racial disparities in health outcomes?
• Better: How do primary care doctors’ rates of referrals to specialists vary by patient race?
• Better: How do rates of health insurance coverage vary by race?

Feasible

• Do you have data with the variables you need?

  • Start with the data—find an interesting dataset; structure your question based on the variables it has.

• Is it simple enough to answer with basic statistics in the next 5 weeks?

Sociological

• Has to do with people…

  • Observations should be people, companies, countries, etc—no plants, animals, cars…

• …living in social systems

  • Question should have a social element
  • Not social: How do stents affect heart attack survival rates?
  • Better: How does the likelihood of stent placement vary by gender identity?

For this class

• Quantitative, not qualitative

  • Not quantitative: How do peoples’ recollections of ads for luxury products vary by socioeconomic status?
  • Better: How do peoples’ ratings of ads for luxury products vary by socioeconomic status?

For this class

• At least two variables
  • How does X affect Y?
  • How does Y vary by X?
  • How does the relationship between X and Y vary by Z?
  • Not enough variables: What proportion of American voters approve of President Biden’s job performance?
  • Better: How does President Biden’s approval rating vary by age among American voters?

Exercise

What are the problems with these research questions? Revise them to improve them.

1. What factors drive gender differences in choice of major?
2. How does susceptibility to migraine headaches vary by age?

Wrapping up computing workflow installation

• Installation instructions on course webpage at Computing -> Setup
• For step 8 (testing it out): instructions for cloning a repo at Computing -> Cloning and committing

For next time

• Next lecture: Studies and sampling (the rest of what you need to know for the proposal)
• Read IMS chapter 2
• Look over instructions for the proposal (Project -> Component 1: Proposal); start looking at data sets and thinking about research questions (bring drafts for peer review Thursday)